One of these things is like the other: Comparisons between treatment-seeking adolescents and clinical trial participants

One of the cornerstones of ED research–of pretty much all medical research, when you get right down to it–is the clinical trial. You take a group of people with a particular illness, give half of them the treatment and give the other half a placebo or no treatment at all. Then, you compare them. Did the treatment group get better? Did they fare significantly better than the control group? What were the factors that affected outcome?

This approach relies upon the fact that the people completing clinical trials for their illness are representative of most people with that condition. People who ask to participate in clinical research might not be your “average” patient. If that is the case, then the treatments that work in a clinical trial might not work for others with the illness. They might have different resources available to them. Their illness or its presentation or symptoms might be different. The treatment may be more or less beneficial. Or even harmful. These types of differences are known as selection bias.

Compare and (don’t) despair

The only way to check if these two groups are the same is to actually look and measure and compare. Thankfully, scientists who study EDs like to compare things almost as much as we sufferers do. Researchers at the University of Chicago compared 214 adolescents who were presenting to their clinic, which does both research and treatment (Stiles-Shields, et al., 2013). Of these, 102 teens participated in research trials and 112 sought standard clinical treatment. For this analysis, the researchers excluded people who were referred to research trials from the clinic, since this might confuse the data. You can see the selection scheme below:

Adapted from Stiles-Shields, et al. 2013. European Eating Disorders Review. DOI: 10.1002/erv.2228

Adapted from Stiles-Shields, et al. 2013. European Eating Disorders Review. DOI: 10.1002/erv.2228

So how did the groups compare? All in all, they were pretty similar. The two groups were identical in terms of average BMI, age, gender, ethnicity, family status (intact vs. not), and presence of binge eating.

One significant difference was that people receiving standard treatment were much more likely to have a diagnosis of EDNOS than clinical trial participants. This is explained by the fact that the clinical trials performed by the University of Chicago generally used the “strict” DSM diagnosis of anorexia and bulimia, though not always. It’s another reason to revamp the EDNOS diagnosis: it will help give us better research data if we can group the ED diagnoses better.

Other differences between the groups were related to family income. Those seeking research treatment had significantly lower family income than those looking for regular treatment (though the income was still fairly high, even in the research group). Again, this makes sense because treatment as part of a clinical trial is free of charge, which can attract families with fewer resources and/or insurance coverage.

The clinical trial participants also had slightly higher scores on the Eating Disorder Examination, which measures ED thoughts and behaviors, greater signs of depression, greater rates of purging behaviors, and lower rates of co-occurring psychiatric disorders.

The study did have some significant limitations. For one, it only looked at people presenting to one particular ED clinic. Other people looking for treatment or participating in research studies at different locations might well have different characteristics. Most people don’t seek care at specialized ED clinics. In fact, the option to use an ED clinic actually isn’t available to many sufferers. These results also might not be representative of the ED population as a whole, since relatively few sufferers receive or seek treatment for their disorder.

The researchers concluded:

In summary, the findings of the present study suggest that the assumption that research samples of adolescents with EDs are less severe than clinic samples is not supported. Future research with improved sampling techniques should examine other psychological factors, such as anxiety, and treatment outcome in relation to research and clinic samples.

Twitter Digg Delicious Stumbleupon Technorati Facebook Email

Comments are closed.